A Review of 2D Lidar SLAM Research

Abstract

Two-dimensional (2D) simultaneous localization and mapping (SLAM) is a key technology for intelligent indoor robots. By using a map generated via SLAM, the robot can navigate and perform specific tasks. This paper reviews the progress of 2D Lidar SLAM algorithms based on four principles: filter-based SLAM, matching-based SLAM, graph optimization-based SLAM, and deep learning-based SLAM, highlighting their advantages, disadvantages, and applicability. Additionally, two key research topics in 2D Lidar SLAM are presented: solutions for dynamic objects during mapping and the fusion of 2D Lidar and vision data. Finally, the development trends of 2D SLAM are discussed.